Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
A navegação consulta e descarregamento dos títulos inseridos nas Bibliotecas Digitais UC Digitalis,
UC Pombalina e UC Impactum, pressupõem a aceitação plena e sem reservas dos Termos e
Condições de Uso destas Bibliotecas Digitais, disponíveis em https://digitalis.uc.pt/pt-pt/termos.
Conforme exposto nos referidos Termos e Condições de Uso, o descarregamento de títulos de
acesso restrito requer uma licença válida de autorização devendo o utilizador aceder ao(s)
documento(s) a partir de um endereço de IP da instituição detentora da supramencionada licença.
Ao utilizador é apenas permitido o descarregamento para uso pessoal, pelo que o emprego do(s)
título(s) descarregado(s) para outro fim, designadamente comercial, carece de autorização do
respetivo autor ou editor da obra.
Na medida em que todas as obras da UC Digitalis se encontram protegidas pelo Código do Direito
de Autor e Direitos Conexos e demais legislação aplicável, toda a cópia, parcial ou total, deste
documento, nos casos em que é legalmente admitida, deverá conter ou fazer-se acompanhar por
este aviso.
Multi-criteria sustainability classification of dairy farms in a portuguese region
Autor(es): Silva, Sandra; Alçada-Almeida, Luís; Dias, Luís C.
Publicado por: Imprensa da Universidade de Coimbra
URLpersistente: URI:http://hdl.handle.net/10316.2/38161
DOI: DOI:http://dx.doi.org/10.14195/978-989-26-1039-9_15
Accessed : 30-Jun-2022 09:01:54
digitalis.uc.ptpombalina.uc.pt
ASSESSMENTMETHODOLOGIESENERGY, MOBILITY AND OTHER REAL WORLD APPLICATION
PEDRO GODINHO JOANA DIASEDITORS
IMPRENSA DA UNIVERSIDADE DE COIMBRACOIMBRA UNIVERSITY PRESS
mulTi-CriTEria SuSTaiNabiliTy ClaSSifiCaTiON
Of Dairy farmS iN a POrTuguESE rEgiON
Sandra Silva1,2, Luís Alçada-Almeida2,3 and Luís C. Dias2,3
Abstract: This paper presents a Multi-criteria Decision Aid assessment of dairy farms
sustainability in the Entre Douro e Minho region in Portugal. Sustainability is a
multidimensional concept encompassing often conflicting economic, social and
environmental points of view, potentially involving various stakeholders and
decision makers. Seven environmental and seven socio-economic criteria were
considered, aided by experts from different fields of knowledge. We used the
multi-criteria classification method ELECTRE TRI to sort dairy farms into three
categories, where socio-economic and environment viability were studied sepa-
rately. The results were then combined to summarize the socio-economic and
environmental viability according to four categories: Highly Sustainable, Moderately
Sustainable, Marginally Sustainable and Not Sustainable. We used a Geographic
Information System for better visualization and analysis of the spatial behavior
of the classification obtained.
Keywords: Multi-criteria Decision Aiding; ELECTRE TRI; Geographic Information
Systems; Agricultural Sustainability.
1 Polytechnic Institute of Viana do Castelo2 INESC - Coimbra, Institute for Systems Engineering and Computers3 Faculty of Economics, University CoimbraE-mails: [email protected], [email protected], [email protected]
http://dx.doi.org/10.14195/978-989-26-1039-9_15
344
1. Introduction
The concept of sustainability is defined in the Brundtland Report
(UN, 1987) as “development which meets the needs of the present
without compromising the ability of future generations to meet their
own needs” arguing that sustainability is based on three distinct pil-
lars: economic, environmental and social. Agricultural sustainability
enhances the quality of life of the population through strengthening
the economy, preserving cultural heritage, protecting natural resources,
and increasing food security. The principles of sustainable develop-
ment thus require that multiple dimensions are taken into account in
sustainability assessment.
Multi-Criteria Decision Aiding (MCDA) methods are suitable for
allowing the consideration of various dimensions associated with an as-
sessment, involving decision makers and potentially other stakeholders
in the evaluation process. This work presents an MCDA methodology for
the classification of dairy farms, according to their sustainability, and
presents an application to the Entre Douro e Minho (EDM) Region, in
Portugal. Initially, environmental concerns and socio-economic concerns
are classified independently. The results are later combined into a single
classification for each farm: Highly Sustainable, Moderately Sustainable,
Marginally Sustainable or Not Sustainable. To perform a classification
the ELECTRE TRI MCDA method (Yu, 1992; Dias et al. 2002; Figueira et
al. 2010) has been selected. A Geographic Information System (GIS) was
used to present the results.
2. Agricultural Sustainability and Multi-criteria Decision Aiding
Agricultural sustainability requires monitoring the agricultural activities
in order to prevent unsustainable behaviors and attitudes. It is necessary
to integrate environmental, social and economic factors in the analysis,
so that all actors involved can coordinate efforts in the quest for sustain-
ability throughout all levels of the agricultural system.
345
The way these dimensions are considered in the evaluation of sus-
tainability varies among different authors. Some authors assume that an
agricultural system is sustainable when the “trade-offs” between the ob-
jectives considered for evaluation of its performances (economic, social
and environmental) reach acceptable values for the society as a whole
(Becker, 1997; Hediger, 1999; Stoorvogel et al., 2004). This approach to
agricultural sustainability becomes possible, while using aggregation
methods to cover the three dimensions mentioned, by aggregating the
set of criteria or indicators into a multidimensional simple or composite
indicator. This approach has been used by several authors as Stockle et
al. (1994), Andreoli and Tellarini (2000), Rigby et al. (2001), van Calker
et al. (2005), Sydorovych and Wossink (2008) and Dillon et al. (2010).
However, the quantification of sustainable agriculture through a set
of indicators is difficult to interpret and often criticized (Hansen, 1996;
Munda, 2005; Gómez-Limón and Riesco, 2009), especially for:
1. the subjectivity of these methods (the choice of functional forms for
the aggregation and weighting of the individual indicators);
2. the compensability usually aggregates the different dimensions of
sustainability attributes (additive aggregation approaches) in spite of
their theoretical incommensurability;
3. not facilitating the process of establishing cause-effect relationships
between diverse system properties and therefore not providing mech-
anisms for diagnosing causes of unsustainability, or for evaluating
effects of proposed interventions.
Taking into account the characteristics that are desired in this kind
of evaluation, we consider that not all aggregation methods would be
appropriate. For example, the requirement of non-compensation seems
to be essential to avoid that a bad performance in one criterion would be
cancelled by a good performance in another criterion. Thus, it was consid-
ered more appropriate to use non-compensatory methods based on the
outranking concept. The outranking approach seemed more appropriate
because, although it is less familiar to decision makers (compared to a
weighted sum) and it requires more configuration parameters, it presents
the following advantages:
346
– it is not compensatory and therefore compatible with the requirements
demanded by the concept of sustainability;
– it introduces the relation of incomparability, useful to account for sit-
uations in which the decision maker and/or the analyst are not able to
compare two actions, which can happen in the sustainability analysis;
– it enables the usage of incomplete value information and, for exam-
ple, judgments on ordinal measurement scale. In the classification of
sustainability, the type of information that exists may be quite diverse
as it comes from various fields (economic, social and environmental)
with different scales and measures.
The sustainability classification of dairy farms has been framed as a
sorting problem (Roy, 1985), because there is no need to compare the
farms among themselves (furthermore, the high number of farms - 1705
- is not adequate for that purpose), but there is need to compare the
farms to predefined targets. ELECTRE TRI, which is a widely used MCDA
Outranking Method, has been chosen because, as a multi-criteria sorting
method, it classifies alternatives (dairy farms) according to preference-
-ordered categories.
The use of multi-criteria methods to assess and determine the sustain-
ability of agriculture is still scarce. Raju el al. (2000) used ELECTRE TRI
and other MCDA techniques for sustainable water resources planning.
Antunes et al. (2011) present the development of a participatory multi-
-criteria analysis process using AHP (Analytic Hierarchy Process) and
SMCE (Social Multi-Criteria Evaluation) to evaluate irrigation management
alternatives. van Calker et al. (2006) applieed MAUT (Multi-Attribute Utility
Theory) to assess economic, social and ecological sustainability of Dutch
dairy farms. Dantsis et al. (2010) applied MAVT (Multi-Attribute Value
Theory) to evaluate and compare the level of sustainability of farms in
two Greek regions. Gómez-Limón and Riesco (2009) and Santos (2011)
applied AHP to the evaluation of agricultural sustainability. Lebacq et
al. (2013) review typologies of sustainability indicators that have been
developed at the farm level and present a case study to help users to
select a set of indicators.
347
3. Outranking Method - ELECTRE TRI
The outranking approach is based on pairwise comparisons between
potential actions or alternatives using a binary, so-called, outranking
relation (Roy, 1996): one action outranks another action if the former is
considered “not worse than” (“at least as good as”) the latter. ELECTRE
methods comprise the construction of one or several outranking relation(s)
comparing pairs of actions in a comprehensive way, followed by an ex-
ploitation procedure (Figueira et al., 2005).
The ELECTRE TRI method (Yu, 1992) provides a complete sorting of
the alternatives of a set of alternatices A into two or more ordered cate-
gories 𝐶𝑖,𝑖=1,…,𝑘, where 𝐶1 is the worst (least desirable) category and 𝐶k
the best one. In order to define the categories, ELECTRE TRI uses some
reference alternatives (reference profiles) 𝑏𝑖, 𝑖=1,…,𝑘−1, which can be real
or hypothetical alternatives. Each reference profile 𝑏𝑖 is simultaneously
the upper bound of category 𝐶𝑖 and the lower bound of category 𝐶𝑖+1.
The assignment of each alternative 𝑎 𝜖 𝐴 to a category is done by compar-
ing it to the reference profiles. Among different ELECTRE TRI variants,
we use the pessimistic (pseudo-conjunctive) procedure (Figueira et al.
2010). Each alternative is assigned to the highest category for which that
alternative outranks its lower bound in the reference profile: a belongs
to C1 if it does not outrank b1, it belongs to C2 if it outranks b1 but it
does not outrank b2, etc.
ELECTRE TRI requires several parameters such as preference (𝑝𝑗 ), indif-
ference (𝑞𝑗 ) and veto (𝑣𝑗 ) thresholds, and weight-importance coefficients
(𝑤𝑗 ). Weight-importance expresses the relative importance of the criteria;
indifference is the largest difference in performance, for a factor, that
may be considered insignificant; preference is the smallest difference in
performance constituting a clear advantage and the veto threshold indi-
cates a difference in performance (discordance) so large that it vetoes
an outranking, even if all other criteria agreed to it.
Each outranking relation is established in four steps: criteria concor-
dance indexes and global concordance indexes calculation; discordance
index calculation; credibility degree calculation; and establishing the
348
outranking relation through the cutting level (λ ∈ [0.5, 1]), that is defined
as the lowest degree of credibility required to assign an alternative to a
specific category. For further details on the algorithm and concepts of this
method, see: Yu (1992), Roy and Bouyssou (1993), Roy (1996), Mousseau
and Dias (2004), and Figueira et al. (2005, 2010).
4. Case Study
4.1. Characterization of Dairy Farms
The study area is corresponds to the Primary Dairy Basin in the EDM
Region located in the Northwest Region of Portugal and it consists of 10
counties: Viana do Castelo, Barcelos, Esposende, Póvoa de Varzim, Vila
Nova de Famalicão, Vila do Conde, Santo Tirso, Trofa, Matosinhos and Maia.
The 1705 dairy farms analyzed in this study are mainly located across
the region of Barcelos (35% of total), Vila do Conde (21%), Póvoa de Varzim
(12%) and Vila Nova de Famalicão (9%). Together, these farms comprise
a total of 106,958 animals. Barcelos (32% of the animals), Vila do Conde
(25%) and Póvoa do Varzim (14%) are the counties with more animals.
The total area of farms in the study measures 20.331,00 hectares
(ha), of which 17.318 ha are Land Area Under Agricultural Production
(LAUAP). The average physical dimension of the total area of each dairy
farms is 12 ha.
4.2. Criteria Selection
Many authors have analyzed the criteria (variables) that explain the
processes of transformation of agricultural production structures. Although
their conclusions do not coincide necessarily, they generally reinforce
three factors that are unanimously considered influential in the possi-
ble paths of structural adjustment (expansion or early abandonment of
agricultural activity): the political and economic factors, the producer
349
(or farmer) and his family, and also the adaptation to environmental
requirements (Santos, 2011).
Sustainable livestock systems should indeed be environmentally friend-
ly, economically viable for farmers, and socially acceptable, notably for
animal welfare. In this work, the criteria are divided into two dimen-
sions: socio-economic and environmental. The definition and evaluation
of performance criteria involved two experts from Environmental and
Zoo-technic fields for the environmental dimension and one expert from
Agro-Economic field for the socio-economic dimension. In defining the
criteria, the experts took into account the new rules for bovine licensing,
the Code of Good Agricultural Practices, Manual Fertilization of crops and
Water Act, and the European legislation and recommendations.
Our study focuses on data-driven assessments carried out at farm level,
from available databases. For the construction of the criteria, the data-
base obtained from the survey developed during the implementation of
the Basin Plan Land Dairy Primary Entre Douro and Minho (POBLPEDM,
2007) was considered, which is the most recent and comprehensive survey
developed in the region. The survey covers some environmental, social
and economic issues of dairy farms; but, not having been developed with
the aim of evaluating dairy farms sustainability, it does not have all the
information that would be ideal to perform a complete economic and
environmental analysis. Financial and economic data as net revenue and
equity to debt are not available. Still, we consider (along with the experts
consulted) that, despite not include the ideal set of criteria, it would be
sufficient to illustrate the methodology developed in this case study.
4.2.1. Socio-Economic Criteria
For an analysis of farmers, Ondersteijn et al. (2003) points out in
the first place the management capabilities of the producer, saying that
they rely on a set of personal characteristics such as their influences,
motivations, skills or biographical factors, age (SE1 criterion in Table 1),
time dedication to work (evaluating the situation of professional activity
– SE2 criterion) and also the professional development of the producers
(indicating their interest in updating and acquiring new knowledge to
350
apply to their operation – SE3 criterion). It is the producer who makes
substantive decisions such as those related to the production system
investments. The characterization of farmers and the important role that
their families play in the management and the activities of the project
are very important.
The importance of family is related to the family tradition in this busi-
ness, considering the involvement of its members and the consequent
existence of successors who demonstrate interest and dedication to farm-
ing (SE4 criterion). Ochoa et al. (2007) discuss the importance of the
succession of generations in agriculture in order to maintain the activity.
Furthermore, the farmer´s (and their family) income earned out of their
dairy farming activity indicates dependence and consequently a need for
business continuity (SE5 criterion).
Since dairy farming is still a particularly demanding activity in terms of
investment in infrastructures and facilities, machinery and effective force,
there can be an obstacle to maintain the motivation for new investments
(SE6 criterion). Finally, it is important analyse the future perspectives
(shutdown, decrease, relocate, stabilize and increase) of the producer
regarding growth and business continuity (SE7 criterion).
The seven socio-economic criteria established are summarized in Table
1 (for more details see Silva, 2015).
Table 1: Socio-Economic Criteria and their Objectives.
ID Socio-Economic Criteria Objective
SE1 Producer´s Age Minimize
SE2 Professional Dedication (%) Maximize
SE3 Professional Development Maximize
SE4 Successors Maximize
SE5 External Income Minimize
SE6 Investment Strategies Maximize
SE7 Future Perspectives Maximize
4.2.2. Environmental Criteria
In the more intensive systems that dominate dairy production, the main
environmental issues are nutrient contamination of soil, groundwater and
351
air pollution. The main impact is on soil integrity which is affected by
increased use of fertilizers, feed additives and the more concentrated use
of waste products such as manure. Other impacts derive from the pollution
of groundwater with nitrates, pesticides and contaminated surface water.
The impact of dairying on the atmosphere arises from de-nitrification,
the production of methane, ammonia volatilization and carbon dioxide.
Consumers and society in general have forced the introduction of new
attitudes with regard to questions of dairy production in particular, such
as animal welfare, environmental impact and traceability and food health
and safety regulations.
Thus, in the environmental dimension, seven criteria are considered
that are defined on the one hand, in accordance with legal requirements
and, on the other hand, with good practice to be implemented, referring
to the cross-compliance, i.e. compliance with standards such as envi-
ronmental health and animal-well-being, maintaining the land in good
agricultural and environmental conditions, and respecting the rules on
public health and pollution. These are summarized in Table 2.
Table 2: Environmental Criteria and their Objectives.
ID Environmental Criteria Objective
E1 Storage Capacity of Manure Maximize
E2 Number of Livestock units per area of manure application Minimize
E3 Excess on nitrogen balance Minimize
E4 Total Production of Greenhouse Gases Minimize
E5 Percentage of Storage Structures Near Water Lines Minimize
E6 Percentage of Individualized Collection of Rainwater Maximize
E7 Animal Well-being Maximize
The first criterion (E1) concerns the farm’s capacity to store manure
produced by its animals and is based on the total storage capacity of
effluents (𝑚3) and the number of standard livestock units (LU), according
to rules for licensing bovine farming. The E2 criterion takes into account
the amounts allowed for the incorporation of organic nitrogen in the soil
in terms of LU/LAUAP, according the Wastewater Management Plan. The
E3 criterion determines the excessive incorporation of nitrogen excreted
352
in relation to the forage area. The E4 criterion defines the production
of greenhouse gases acoording the production of milk of each dairy
farm. The E5 criterion addresses the risk of contaminating water lines
and consists of the ratio between the number of storage points less than
25 meters away from the water lines and the total number of existing
storages on the farm. The E6 criterion addresses the use of water and
consists of the proportion of the farms’ storage facilities that have
individualized collection of rainwater, that increases the effectiveness
of storage capacity, provides fertilizers of better quality, reuses water,
and thereby reduces the water burden. The E7 criterion concerns animal
well-being, assessed by considering the ratio between the covered area
(stall) and the number of cows on the farm. For more details on these
criteria see Silva (2015).
4.3. Application of the ELECTRE TRI Method
ELECTRE TRI is a suitable MCDA method in sorting problems, as
addressed in this case study of dairy farms classification according to
their environmental and socio-economic viability. In a classification of
dairy farms in which social, economic and environmental aspects are at
stake, it is often important to use known standards (references profiles
defined in legislation) to define various categories.
ELECTRE TRI assigns a set of alternatives to a group of predefined cat-
egories, considering multiple criteria. In this work, the set of alternatives
are dairy farms and the predefined categories consist of three viability
levels: Not Viable (Category 1), Barely Viable (Category 2) and Viable
(Category 3). We study the environmental and socio-economic viability
separately due to the distinct nature of these two areas, and then derive
a broader perspective on sustainability.
The ELECTRE TRI method performs the assignment of each alternative
to a category, based on how its performances compare with the reference
profiles. We define two reference profiles 𝑏1 and 𝑏2, where 𝑏2 indicates
the minimum performance standards that a farm should have in order
353
to be placed in Category 3 and 𝑏1 indicates the minimum performance
standards in order to be in Category 2. The profiles (Table 3) have been
defined by experts.
Table 3: References Profiles (𝑏1 and 𝑏2) for the Environmental and Socio-Economic classification using the ELECTRE TRI method.
ENVIRONMENTAL SOCIO-ECONOMICE1 E2 E2 E4 E5 E6 E7 SE1 SE2 SE3 SE4 SE5 SE6 SE7
𝑏2 1.0 1.0 0 1000 0.25 0.5 5.0 𝑏1 50 4 3 4 2 1 4
𝑏1 0.8 1.3 300 4000 0.50 0.2 2.5 𝑏1 65 2 0 2 4 0 1
Table 4: Environmental and Socio-Economic Thresholds of the ELECTRE TRI method: indifference (𝑞𝑗 ), preference (𝑝𝑗 ), and veto (𝑣𝑗 ) thresholds,
and weight-importance coefficients (𝑤𝑗 ).
ENVIRONMENTALE1 E2 E3 E4 E5 E6 E7
𝑞𝑗 0.10 0.10 100 300 0.05 0.10 1
𝑝𝑗 0.20 0.20 150 800 0.10 0.20 2
𝑣𝑗 0.50 0.50 500 6000 0.55 0.45 3
𝑤𝑗 0.20 0.20 0.20 0.05 0.15 0.05 0.15
SOCIO-ECONOMICSE1 SE2 SE3 SE4 SE5 SE6 SE7
𝑞𝑗 10 1 1 1 1 0 1
𝑝𝑗 15 2 2 2 2 0.5 2
𝑣𝑗 20 4 3 4 3 1 3
𝑤𝑗 0.2 0.1 0.1 0.2 0.1 0.15 0.15
Taking into account the inaccuracies and uncertainties in the perfor-
mances of the alternatives, thresholds of indifference (𝑞𝑗 ) and preference
(𝑝𝑗 ) were defined for each criterion, in order to better model the prefer-
ences of the decision makers. The veto threshold (𝑣𝑗 ) that is used in the
discordance test was also determined for each criterion, as well as the
weight (𝑤𝑗 ) of each criterion that is used in the concordance test (Table
4). We calculated the partial concordance index for each of the criteria
and then the comprehensive concordance indices of global agreement
that indicates how much is the performance of the alternatives consis-
tent with “𝑎 outranks 𝑏𝑖”, i.e. 𝑎𝑆𝑏𝑖. Most of the criteria (sum of weights
354
above the cutting level) should be in favor of this statement. Next, we
calculated the discordance indexes for each criterion, representing the
statement opposed to 𝑎𝑆𝑏𝑖.
Finally, the credibility index (𝜎(𝑎,𝑏𝑖)) was calculated, which determ-
ines how credible the statement 𝑎𝑆𝑏𝑖 is according to the comprehensive
concordance and discordance indexes. For the classification procedure
it was also necessary to define a cutting level (λ), in this case λ = 0.6, to
determine the binary relationship between the alternatives and reference
profiles, such that: 𝑎𝑆𝑏𝑖 ⇔ 𝜎(𝑎,𝑏𝑖)≥λ. The cutting level indicates that 60%
of the criteria (considering their weights) have to agree in order to assign
an alternative to a specific category.
Individual meetings were held with each expert, where the method
was presented and the values of its parameters were elicited.
5. Results
A Web Multicriteria Spatial Decision Support System (Web MC-SDSS)
named ELECTRE TRI in ArcGIS has been applied in this work (for details
see Silva et al., 2014) to classify the environmental and socio-economic
viability of dairy farms, obtaining the results presented in Figure 1 and
the map with the viability classification.
Concerning environmental viability, we conclude that the majority (60%)
of dairy farms are classified as Environmentally Not Viable (Category 1).
On the other hand, the assessment of the Socio-economic Viability, con-
sidering the defined criteria, indicates that only a small minority (3%) is
classified as Not Viable. The analysis of Figure 1 shows that environmental
issues are those that mostly affect the viability of dairy farms in the EDM
Region, for non-compliance with environmental requirements. Another
important aspect is the noticeable difference between the classifications
obtained in the evaluation of Environmental and Socio-economic Viability
of farms. Let us note that these two dimensions are not completely inde-
pendent: a farm could improve its environmental condition by making
more investments, at the cost of hindering its economic condition, unless
355
subsidies or other incentives were made available (unfortunately in the
current situation of this region such investments are not likely). Thus,
the classification Environmentally Not Viable should be understood as
meaning that a farm is hardly viable due to its environmental condition.
Figure 1: Number of farms classified in each category.
Table 5 compares the classifications obtained according to Environmental
Viability and the Socio-economic Viability. Only 2% (30) of the dairy farms
are classified as Environmentally Not Viable and Socio-economically Not
Viable; 16% (271) of the farms are classified as Barely Viable, considered
to be environmental and Socio-economic Barely Viable simultaneous-
ly. Only 3% (48) of the farms are classified as Viable according to the
Environmental and the Socio-economic perspectives. The largest group,
with 44% (754) of the total number dairy farms, is classified as both
Socioeconomically Barely Viable and Environmentally Not Viable.
Table 5: Cross tabulation of the classification of Environmental and Socio-Economic Viability.
Socio-Economic ViabilityNot Viable Barely Viable Viable
Environmental Not Viable 30 754 234 1018
Viability Barely Viable 13 271 69 353
Viable 8 278 48 334
51 1303 351 1705
After obtaining these results, the experts were again involved in a
new discussion and analysis, considering that in order to conduct the
356
evaluation of a farm as to the overall Sustainability, it would make
sense to reclassify (or sub-divide) the categories obtained in the study
of Environmental and Socioeconomic Viability into the following new
set of categories:
Highly Sustainable - dairy farms classified as Viable on Socio-Economic
and Environmental Viability, simultaneously;
Moderately Sustainable - dairy farms classified as Barely Viable in
Environmental Viability and Socio-Economic Viability in Viable or vice
versa;
Marginally Sustainable - dairy farms classified as Barely Viable in
Environmental and Socioeconomic Viability;
Not Sustainable - all other.
This reclassification yields the results depicted in Figure 2.
Figure 2: Number and percentage of dairy farms classified in each of four categories of Sustainability
Figure 3 presents the corresponding kernel density maps for a better
visualization of the distribution of the density of each of these four cate-
gories. The Kernel Density (available in ArcGIS software) consists in an
interpolation which is intended to generate a continuous surface through
sample points (in this case, exploits), which is determined by number of
farms classified in each category in relation to the area (km2).
357
Figure 3: Density of distribution of dairy farms according to the new (aggregated) perspective of Sustainability
Analyzing briefly the spatial distribution of the sustainability classi-
fication, we can highlight the following: about 90% of the dairy farms
in the county of Vila do Conde are classified as Not Sustainable, as
well as 88% of the dairy farms in Póvoa de Varzim and 82% of those in
Esposende – which are very high percentages. The counties of Maia,
Santo Tirso and Barcelos are those with better classifications: there is
a higher percentage of farms classified as Highly Sustainable and the
lowest percentage of dairy farms classified as Not Sustainable (but
still around 40%). The dairy farms classified as Not Sustainable, have
in general a high animal density and reduced dimensions, and have
difficulties in growth due to their proximity to urban areas. In turn, in
dairy farms classified as High Sustainable there is a better fit between
the number of animals and their dimension but especially they have
improved environmental conditions.
358
6. Sensitivity Analysis
Setting the required input data for the ELECTRE TRI method (weights,
thresholds of indifference and preference, veto, reference profiles and
cutting level) is a complex and ambiguous task because of the associated
uncertainties. To check how changes in the parameters influence the re-
sults, we performed a sensitivity analysis, considering different scenarios,
in order to verify the robustness of the results.
Two scenarios were defined, in order to develop a sensitivity analysis
in the study of Environmental and Socio-economic Viability, which we
will call the Benevolent and Stringent scenarios. These scenarios are
characterized by a variation of 5% in the profiles of the categories, in
the preference, indifference and veto thresholds and in the cutting level
(there was no change in the weights in any of the scenarios). We aim at
evaluating changes in the results obtained in these two scenarios when
compared with the classification of originally obtained data, which we
shall call the Original scenario.
The Benevolent Scenario is characterized by changing the parameters
of the ELECTRE TRI method as follows: the cutting level is decreased by
0.03 (becoming 0.57), the reference profiles of the categories increase
5% if a criterion is to be minimized and decrease 5% if a criterion
is to be maximized, and thresholds of preference, indifference and
veto are increased by 5%. This is a less demanding scenario than the
Original one.
In turn, the Stringent Scenario considers the changes referred to in
the following parameters of the ELECTRE TRI method: the cutting level
is increased by 0.03 (becoming 0.63), the reference profiles of the cate-
gories decrease 5% if the criterion is to be minimized, and increase 5% if
a criterion is to be maximized, and thresholds of preference, indifference
and veto are decreased by 5%. This is a more demanding scenario than
the original one.
The results of comparing the three scenarios are summarized in Figure 4.
359
Figure 4: Comparison of the three scenarios regarding the classification of Socio-Economic and Environmental Viability (Category 1 - Not Viable,
Category 2 – Barely Viable, Category 3 - Viable).
In the Stringent Scenario, the percentage change in the socio-economic
assessment is lower than in the environmental assessment. The percent-
age change in each category is equal to or less than the variation caused
in the parameters and, as such, it can therefore be said that despite the
change in the final results this is not very relevant.
In the Benevolent Scenario, it is noted that the number of dairy farms
changing classification in environmental analysis is less than the number
changing in the socio-economic analysis. This means that, even when
requiring less compliance with environmental requirements, no major
changes in the classification of farms are visible. Despite the fact that
there were some changes obtained in the classification, those were not
very significant.
We considered also other Stringent and Benevolent scenarios char-
acterized by a greater variation: 15% in the profiles of the categories,
in the preference, indifference and veto thresholds and cutting level
(becoming 0.51 in Benevolent and 0.69 in Stringent scenarios). In both
new Stringent and Benevolent scenarios the changes in the classification,
in all four sustainability categories, is less than percentage change, not
exceeding 7%, compared with results of Original scenario. In Figure 5,
it is possible compare the number of dairy farms classified in each of
four categories of Sustainability for the 5 scenarios analysed: Stringent
360
(5% and 15%), Original and Benevolent (5% and 15%) scenarios. These
different scenarios considered support the stability of the results obtained
by ELECTRE TRI method.
Figure 5: Number of dairy farms classified in each of four categories of Sustainability by comparing the five scenarios.
7. Discussion
Considering the results, one notes the worrying situation in which
many farms in the EDM Region are found regarding their sustainability.
In this region, the economic component contributes greatly to the sustain-
ability of farms, but the breach of environmental requirements is having
a greater influence on their sustainability. There are reasons, natural and
historical, which tend to hinder the resolution of environmental problems,
such as the ownership structure of dairy farms and their dispersion, the
morphological and geological nature of the terrains and the proximity to
urban areas. The solution to this environmental problem is an evolution
of the structure of dairy farms to reduce the intensity of production, but
also for obtaining dimension which will amortize investments in equip-
ment and training. However, the economic framework is itself heavily
influenced by the nature of the support regime (largely price support),
361
technology development and structural change in the production sector
(plus structural change in the upstream and downstream supply chain).
So it is necessary to find diverse environmentally and economically
sustainable solutions, for those farms that are moderately or marginally
sustainable. As for the farms that are not sustainable, although solutions
should be sought to enable the continuation of the dairy farming activity,
a reconversion to a different activity is also an alternative to consider.
For this study, the ELECTRE TRI method seemed appropriate as it sorts
alternatives (dairy farms) by preference-ordered categories. Indeed, in
classification of dairy farms, where various environmental, economic and
social aspects are at stake, it is often important to use known standards
(reference profiles) to define various categories of sustainability. This
requires capability of evaluating each farm in absolute terms, not just in
comparison with other peers, as well as the need to include evaluation
aspects expressed in different units, using any type of scales (including
qualitative). The ELECTRE TRI refuses the possibility of total compensation
between the alternative’s performances on the criteria. Thus it prevents
that a farm with very good performance on some criterion compensates
a very poor performance on another criterion and achieves the best cat-
egory despite that major weakness.
Finally, the sensitivity analysis developed verifies the robustness of
the results of dairy farms sustainability assessment.
8. Conclusions
The complexity of sustainability assessment can be mitigated with the
use of MCDA techniques. In this study, we found that MCDA can be useful
in the evaluation of sustainability that encompasses multiple dimensions.
The Outranking Method ELECTRE TRI proved to be an adequate method,
preventing unwarranted compensation of poor performances, and yielding
a classification of dairy farms on sustainability according to the parame-
ters and values assigned by experts. The integration of GIS capabilities
and MCDA techniques reinforces the advantages of both tools.
362
The main highlights of the methodology followed in this paper are
the following. First, the sustainability assessment is based explicitly on
multiple criteria, rather than trying to find an indicator such as carbon
footprint or a financial ratio that would synthetize environmental and
economic sustainability, respectively. Second, the evaluation is grounded
on a classification method which is parsimonious in terms of the value
judgments it requires, namely not requiring substitution rates among
the criteria, nor pairwise comparisons among a large number of alter-
natives. Although the results are somewhat coarse (when compared to a
full ranking), classification results are familiar and widely accepted by
decision makers (e.g., efficiency labeling). Third, a separate assessment
of sustainability dimensions is made, directing attention to different po-
tential weaknesses of the farms. Nevertheless, these two dimensions of
sustainability can be conveniently summarized into an easy to understand
overall classification, namely to present graphically in maps.
In this work it was also possible to analyze the geographical distribu-
tion of the classification of dairy farms and to realize that the areas which
reflect higher levels of sustainability are those that have an increasing
concentration of farms, whereas areas that reflect the worst levels of
sustainability are those that are subject to abandonment.
The structure, process and application of sustainability assessment can
be extended to other areas besides agriculture. In this decision process,
options were taken in each phase, which leave space for future studies
and the development of new approaches, such as the aggregation of
these or other criteria, with the involvement of more experts with the
application of other(s) multi-criteria method(s), with application for more
updated reinvestigation.
References
Andreoli, M., Tellarini, V., 2000. Farm sustainability evaluation: methodology and practice. Agriculture, Ecosystems and Environment, 77(1): 43-52.
363
Antunes, P., Karadzic, V., Santos R., Beça P., Osann, A., 2011. Participatory multi-criteria analysis of irrigation management alternatives: the case of the Caia irrigation district, Portugal. International Journal of Agricultural Sustainability, 9(2): 334-349.
Becker, B., 1997. Sustainability assessment: a review of values, concepts, and methodological approaches. Consultative Group on International Agricultural Research, The World Bank, Washington, D.C., 63 pp.
Dantsis, T., Douma, C., Giourga, C., Loumou, A., Polychronaki, E.A., 2010. A methodological approach to assess and compare the sustainability level of agricultural plant production systems. Ecological Indicators, 10: 256–263.
Dias, L., Mousseau, V., Figueira, J., Clímaco, J., 2002. An aggregation/disaggregation approach to obtain robust conclusions with ELECTRE TRI. European Journal of Operational Research, 138: 332–348.
Dillon, E. J., Hennessy T., Hynes, S., 2010. Assessing the sustainability of Irish agriculture. International Journal of Agricultural Sustainability, 8(3): 131-147
Figueira, J., Mousseau, V., Roy, B., 2005. Electre Methods. In: Figueira, J.; Greco, S.; Ehrgott, M. Multiple Criteria Decision Analysis: State of The Art Survey. New York: Springer, Cap. 4.
Figueira, J.R., Greco S., Roy B., Słowinski R., 2010. ELECTRE methods: Main features and recent developments. In C. Zopounidis and P. Pardalos (Eds.), Handbook of Multicriteria Analysis, Chapter 4, New York, USA: Springer.
Gómez-Limón, J. A., Riesgo, L., 2009. Alternative Approaches to the construction of a composite indicator of agricultural sustainability: An application to irrigated agriculture in the Duero basin in Spain. Journal of Environmental Management, 90: 3345-3362.
Hansen, J.W., 1996. Is agricultural sustainability a useful concept? Agricultural Systems, 50(1): 117-143.
Hediger, W., 1999. Reconciling ‘weak’ and ‘strong’ sustainability. International Journal of Social Economics, 26:7/8/9, 1120-1143.
Lebacq, L., Baret, P.V., Stilmant, D. 2013. Sustainability indicators for livestock farming. A review. Agronomy for Sustainable Development, 33:311-327.
Mousseau, V., Dias, L., 2004. Valued outranking relations in ELECTRE providing manageable disaggregation procedures. European Journal of Operational Research, 156: 467–482.
Munda, G., 2005. Measuring sustainability: a multi-criterion framework. Environment, Development and Sustainability, 7(1):117-134.
Ochoa, A.M.A, Oliva, V.C, Sáez, C.A, 2007. Explaining farm succession: the impact of farm location and off-farm employment opportunities. Spanish Journal of Agricultural Research, 5(2): 214-225.
Ondersteijn, C.J.M., Giesen, G.W.J., Huirne, R.B.M., 2003. Identification of farmer characteristics and farm strategies explaining changes in environmental management and environmental and economic performance of dairy farms. Agricultural Systems, 78: 31-55.
POBLPEDM, 2007. Plano de Ordenamento da Bacia Leiteira Primária do Entre Douro e Minho. Escola Superior Agrária do Instituto Politécnico de Viana do Castelo, Instituto para o Desenvolvimento Agrário da Região Norte e Universidade do Porto – CIBIO.
Raju, K.S., Duckstein, L., Arondel, C., 2000. Multicriterion Analysis for Sustainable Water Resources Planning: A Case Study in Spain. Water Resources Management, 14: 435–456.
Rigby, D., Woodhouse, P., Young, T. and Burton, M., 2001. Constructing a farm level indicator of sustainable agricultural practice. Ecological Economics, 39(3): 463-478.
Roy, B., 1985. Méthodologie multicritère d’aide à la decision. Economica, Paris.
364
Roy B., Bouyssou, D., 1993. Aide multicritère à la décision: méthodes et cas, Economica, Paris.
Roy, B., 1996. Multicriteria Methodology for Decision Aiding. Kluwer Academic Publishers.
Santos, J.C.S.M., 2011. Viabilidad de las explotaciones lecheras en un escenario de desconexión del apoyo y condicionalidad: Un estudio de caso en el noroeste de Portugal. Ph.D. Dissertation, Universidad Politécnica de Madrid, Escuela Técnica Superior de Ingenieros Agrónomos.
Silva, S., 2015. Sistema de Apoio à Decisão Espacial Multicritéio na localização de centrais de biogás. Tese de Doutoramento em Gestão-Ciência Aplicada à Decisão. Faculdade de Economia, Universidade de Coimbra.
Silva, S., Almeida-Alçada, L., Dias L., 2014. Development of a Web-based Multi-criteria Spatial Decision Support System for the assessment of environmental sustainability of dairy farms, Computers and Electronics in Agriculture, 108: 46–57.
Stockle, C.O., Papendick, R.I., Saxton, K.E., Campbell, G.S., van Evert, F.K., 1994. A framework for evaluating the sustainability of agricultural production systems. American Journal of Alternative Agriculture, 9(1-2): 45-50.
Stoorvogel, J.J., Antle, J.M., Crissman, C.C. and Bowen W., 2004. The tradeoff analysis model: integrated biophysical and economic modeling of agricultural production systems. Agricultural Systems, 80(1): 43-66.
Sydorovych, O., Wossink, A., 2008. The meaning of agricultural sustainability: Evidence from a conjoint choice survey. Agricultural Systems, 98(1): 10-20.
United Nations, 1987. Brundtland Report, Report of the World Commission on Environment and Development: Our Common Future, Oxford University Press.
van Calker, K., Berentsen, P., Romero, C., Giesen, G., Huirne, R., 2006. Development and application of multi‐attribute sustainability function for Dutch dairy farming systems. Ecological Economics, 57: 640‐658.
van Calker, K., Berentsen P., Giesen G., Huirne, R., 2005. Identifying and Ranking Attributes that Determine Sustainability in Dutch Dairy Farming. Agriculture and Human Values, 22: 53-63.
Yu, W., 1992. ELECTRE TRI: Aspects méthodologiques et manuel d’utilisation. Document du LAMSADE No. 74, Université Paris-Dauphine.
Série Investigação
•
Imprensa da Universidade de Coimbra
Coimbra University Press
2015